Original research at the intersection of AI, organizational design, and the infrastructure needed for both. Organized by project.
The Organizational Context Project: research into AI-native memory, context loss, and the data architectures needed to fix it. Kendriala is the reference implementation for the hypotheses developed here.
Foundations, Architectures, and Open Problems for AI-Native Organizational Memory
Argues that context loss is a structural property of CRUD-centric design; that LLMs convert this into a binding constraint; and that solving it requires treating context as a first-class engineering artifact with its own data model, temporal semantics, and measurable quality.
A Design Case Study
Architecture decisions registered as testable hypotheses, a build sequence as research design, and pre-registered failure signatures: written before results exist so they can confirm or embarrass the design when they arrive.